TMP- SSurface2: A Novel Deep Learning-Based Surface Accessibility Predictor for Transmembrane Protein Sequence

Zhang

et al. 2021

Preprint

One of the unique traits of membrane proteins is that a significant fraction of their hydrophobic amino acids is exposed to the hydrophobic core of lipid bilayers rather than being embedded in the protein interior, which is often not explicitly considered in protein structure prediction. Here, we propose a characteristic and predictive quantity, the lipid contact probability (LCP), to describe the likelihood of the amino acids of a given sequence being in direct contact with the acyl chains of lipid molecules. We show that LCP is complementary to solvent accessibility in characterizing the outer surface of membrane proteins, and it can be predicted for any given sequence with a machine learning-based method by utilizing a training dataset extracted from MemProtMD, a database generated from molecular dynamics simulations for the membrane proteins with a known structure. We demonstrate that LCP can be used to systematically improve the prediction accuracy of the contact maps of proteins, as well as to identify membrane-anchored soluble proteins from sequences.

Section: Generation Of the Dataset For The Mcp Model Trainingmentioning

confidence: 99%

Membrane contact probability: an essential and predictive character for the structural and functional studies of membrane proteins

Zhang

et al. 2021

Preprint

J. Chem. Theory Comput.

“…Last, TMP-SSurface2 predicts the RSA for residues placed in a membrane environment rather than a direct exposure classification. Interestingly, the Pearson’s R in the same range as for our independent OPM test set (Figure D) was reported …”

Section: Discussionmentioning

confidence: 65%

“…69 The same predictors but with different relative performances were identified for the OPM test set. On the OPM test set, we also compared TopProperty RSA predictions to TMP-SSurface-2, 70 a recently published method specifically trained on TMPs. Although TopProperty was not trained specifically on TMPs, it performed equivalently or better compared to TMP-SSurface-2.…”

Section: ■ Resultsmentioning

confidence: 99%

“…Other methods have been described for predicting directly or indirectly the membrane exposure of TMP residues, such as LIPS, RHYTHM, TMX, and more recently, TMP-SSurface2 and LCP . From these predictors, LIPS and RHYTHM provide information regarding exposure per α-helix face, obscuring the per-residue signal, and lack information for TMBs; TMX is not available anymore.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

TopProperty: Robust Metaprediction of Transmembrane and Globular Protein Features Using Deep Neural Networks

Mulnaes

Schott‐Verdugo

Koenig

et al. 2021

Transmembrane proteins (TMPs) are critical components of cellular life. However, due to experimental challenges, the number of experimentally resolved TMP structures is severely underrepresented in databases compared to their cellular abundance. Prediction of (per-residue) features such as transmembrane topology, membrane exposure, secondary structure, and solvent accessibility can be a useful starting point for experimental design or protein structure prediction but often requires different computational tools for different features or types of proteins. We present TopProperty, a metapredictor that predicts all of these features for TMPs or globular proteins. TopProperty is trained on datasets without bias toward a high number of sequence homologs, and the predictions are significantly better than the evaluated state-of-theart primary predictors on all quality metrics. TopProperty eliminates the need for protein type-or feature-tailored tools, specifically for TMPs. TopProperty is freely available as a web server and standalone at https://cpclab.uni-duesseldorf.de/topsuite/.

“…This smaller dataset was termed ‘MCP-Small’. To ensure that the training set is relatively sufficient [ 24 , 62 ], the 898 sequences were divided into three subsets: 718 randomly chosen membrane protein sequences formed the training set, 90 membrane protein sequences were used as the test set, and the other 90 membrane protein sequences were used as the validation set. Such a ∼8:1:1 dataset construction ensures a relatively larger training set with reasonable validation and test sets for better convergence, which was adopted and recommended by previous work [ 63 , 64 ].…”

Section: Methodsmentioning

confidence: 99%

Membrane contact probability: An essential and predictive character for the structural and functional studies of membrane proteins

et al. 2022

One of the unique traits of membrane proteins is that a significant fraction of their hydrophobic amino acids is exposed to the hydrophobic core of lipid bilayers rather than being embedded in the protein interior, which is often not explicitly considered in the protein structure and function predictions. Here, we propose a characteristic and predictive quantity, the membrane contact probability (MCP), to describe the likelihood of the amino acids of a given sequence being in direct contact with the acyl chains of lipid molecules. We show that MCP is complementary to solvent accessibility in characterizing the outer surface of membrane proteins, and it can be predicted for any given sequence with a machine learning-based method by utilizing a training dataset extracted from MemProtMD, a database generated from molecular dynamics simulations for the membrane proteins with a known structure. As the first of many potential applications, we demonstrate that MCP can be used to systematically improve the prediction precision of the protein contact maps and structures.